-
Notifications
You must be signed in to change notification settings - Fork 297
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Ingress controller crash if endpoint slice missing service label #1730
Comments
/kind bug |
/assign |
Thanks @mikenorgate for finding the issue. We will take a look and update as soon as we can. |
@mikenorgate, are these logs from ingress-gce? Are you using the inbuilt ingress-gce instance in GKE or running it yourself? |
* Service key may not be possible if EndpointSlice does not have all the expected labels set, namely the service label. This is a valid EPS and should be ignored when indexing. Returning an error causes the store to panic. Instead return empty list of keys and no error, so that the EPS is ignored. Fixes kubernetes#1730
I was using the inbuilt instance when I initially noticed the problem but was unable to find any logs to indicate the cause. I then followed the instructions from here (https://github.com/kubernetes/ingress-gce/tree/master/docs/deploy/gke) to deploy a version where I could monitor the logs |
I am running a GKE cluster (version 1.22.8-gke.201) in which I have setup Submariner (https://submariner.io/) for multi-cluster communication. As part of Submarniners operation it uses endpoint slices to keep track of available endpoints across the clusters.
These endpoint slices are missing a service label which causes a panic in the GCE ingress controller
The text was updated successfully, but these errors were encountered: